1 /* Definitions for data structures and routines for the regular 2 expression library, version 0.12. 3 4 Copyright (C) 1985, 1989, 1990, 1991, 1992, 1993 Free Software Foundation, Inc. 5 6 This program is free software; you can redistribute it and/or modify 7 it under the terms of the GNU General Public License as published by 8 the Free Software Foundation; either version 2, or (at your option) 9 any later version. 10 11 This program is distributed in the hope that it will be useful, 12 but WITHOUT ANY WARRANTY; without even the implied warranty of 13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 14 GNU General Public License for more details. 15 16 You should have received a copy of the GNU General Public License 17 along with this program; if not, write to the Free Software 18 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ 19 20 #ifndef __REGEXP_LIBRARY_H__ 21 #define __REGEXP_LIBRARY_H__ 22 23 /* POSIX says that <sys/types.h> must be included (by the caller) before 24 <regex.h>. */ 25 26 #if defined(VMS) || defined(__MACOS__) 27 /* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it 28 should be there. */ 29 #include <stddef.h> 30 #endif 31 32 33 #undef NLM_EXTERN 34 #ifdef NLM_IMPORT 35 #define NLM_EXTERN NLM_IMPORT 36 #else 37 #define NLM_EXTERN extern 38 #endif 39 40 #ifdef __cplusplus 41 extern "C" { 42 #endif 43 44 45 /* The following bits are used to determine the regexp syntax we 46 recognize. The set/not-set meanings are chosen so that Emacs syntax 47 remains the value 0. The bits are given in alphabetical order, and 48 the definitions shifted by one from the previous bit; thus, when we 49 add or remove a bit, only one other definition need change. */ 50 typedef unsigned reg_syntax_t; 51 52 /* If this bit is not set, then \ inside a bracket expression is literal. 53 If set, then such a \ quotes the following character. */ 54 #define RE_BACKSLASH_ESCAPE_IN_LISTS (1) 55 56 /* If this bit is not set, then + and ? are operators, and \+ and \? are 57 literals. 58 If set, then \+ and \? are operators and + and ? are literals. */ 59 #define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) 60 61 /* If this bit is set, then character classes are supported. They are: 62 [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], 63 [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. 64 If not set, then character classes are not supported. */ 65 #define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) 66 67 /* If this bit is set, then ^ and $ are always anchors (outside bracket 68 expressions, of course). 69 If this bit is not set, then it depends: 70 ^ is an anchor if it is at the beginning of a regular 71 expression or after an open-group or an alternation operator; 72 $ is an anchor if it is at the end of a regular expression, or 73 before a close-group or an alternation operator. 74 75 This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because 76 POSIX draft 11.2 says that * etc. in leading positions is undefined. 77 We already implemented a previous draft which made those constructs 78 invalid, though, so we haven't changed the code back. */ 79 #define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) 80 81 /* If this bit is set, then special characters are always special 82 regardless of where they are in the pattern. 83 If this bit is not set, then special characters are special only in 84 some contexts; otherwise they are ordinary. Specifically, 85 * + ? and intervals are only special when not after the beginning, 86 open-group, or alternation operator. */ 87 #define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) 88 89 /* If this bit is set, then *, +, ?, and { cannot be first in an re or 90 immediately after an alternation or begin-group operator. */ 91 #define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) 92 93 /* If this bit is set, then . matches newline. 94 If not set, then it doesn't. */ 95 #define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) 96 97 /* If this bit is set, then . doesn't match NUL. 98 If not set, then it does. */ 99 #define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) 100 101 /* If this bit is set, nonmatching lists [^...] do not match newline. 102 If not set, they do. */ 103 #define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) 104 105 /* If this bit is set, either \{...\} or {...} defines an 106 interval, depending on RE_NO_BK_BRACES. 107 If not set, \{, \}, {, and } are literals. */ 108 #define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) 109 110 /* If this bit is set, +, ? and | aren't recognized as operators. 111 If not set, they are. */ 112 #define RE_LIMITED_OPS (RE_INTERVALS << 1) 113 114 /* If this bit is set, newline is an alternation operator. 115 If not set, newline is literal. */ 116 #define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) 117 118 /* If this bit is set, then `{...}' defines an interval, and \{ and \} 119 are literals. 120 If not set, then `\{...\}' defines an interval. */ 121 #define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) 122 123 /* If this bit is set, (...) defines a group, and \( and \) are literals. 124 If not set, \(...\) defines a group, and ( and ) are literals. */ 125 #define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) 126 127 /* If this bit is set, then \<digit> matches <digit>. 128 If not set, then \<digit> is a back-reference. */ 129 #define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) 130 131 /* If this bit is set, then | is an alternation operator, and \| is literal. 132 If not set, then \| is an alternation operator, and | is literal. */ 133 #define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) 134 135 /* If this bit is set, then an ending range point collating higher 136 than the starting range point, as in [z-a], is invalid. 137 If not set, then when ending range point collates higher than the 138 starting range point, the range is ignored. */ 139 #define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) 140 141 /* If this bit is set, then an unmatched ) is ordinary. 142 If not set, then an unmatched ) is invalid. */ 143 #define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) 144 145 /* This global variable defines the particular regexp syntax to use (for 146 some interfaces). When a regexp is compiled, the syntax used is 147 stored in the pattern buffer, so changing this does not affect 148 already-compiled regexps. */ 149 NLM_EXTERN reg_syntax_t re_syntax_options; 150 151 /* Define combinations of the above bits for the standard possibilities. 152 (The [[[ comments delimit what gets put into the Texinfo file, so 153 don't delete them!) */ 154 /* [[[begin syntaxes]]] */ 155 #define RE_SYNTAX_EMACS 0 156 157 #define RE_SYNTAX_AWK \ 158 (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ 159 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 160 | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ 161 | RE_UNMATCHED_RIGHT_PAREN_ORD) 162 163 #define RE_SYNTAX_POSIX_AWK \ 164 (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS) 165 166 #define RE_SYNTAX_GREP \ 167 (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ 168 | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ 169 | RE_NEWLINE_ALT) 170 171 #define RE_SYNTAX_EGREP \ 172 (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ 173 | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ 174 | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ 175 | RE_NO_BK_VBAR) 176 177 #define RE_SYNTAX_POSIX_EGREP \ 178 (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES) 179 180 /* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ 181 #define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC 182 183 #define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC 184 185 /* Syntax bits common to both basic and extended POSIX regex syntax. */ 186 #define _RE_SYNTAX_POSIX_COMMON \ 187 (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ 188 | RE_INTERVALS | RE_NO_EMPTY_RANGES) 189 190 #define RE_SYNTAX_POSIX_BASIC \ 191 (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) 192 193 /* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes 194 RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this 195 isn't minimal, since other operators, such as \`, aren't disabled. */ 196 #define RE_SYNTAX_POSIX_MINIMAL_BASIC \ 197 (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) 198 199 #define RE_SYNTAX_POSIX_EXTENDED \ 200 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 201 | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ 202 | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ 203 | RE_UNMATCHED_RIGHT_PAREN_ORD) 204 205 /* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS 206 replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */ 207 #define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ 208 (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ 209 | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ 210 | RE_NO_BK_PARENS | RE_NO_BK_REFS \ 211 | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) 212 /* [[[end syntaxes]]] */ 213 214 215 /* Maximum number of duplicates an interval can allow. Some systems 216 (erroneously) define this in other header files, but we want our 217 value, so remove any previous define. */ 218 #ifdef RE_DUP_MAX 219 #undef RE_DUP_MAX 220 #endif 221 #define RE_DUP_MAX ((1 << 15) - 1) 222 223 224 /* POSIX `cflags' bits (i.e., information for `regcomp'). */ 225 226 /* If this bit is set, then use extended regular expression syntax. 227 If not set, then use basic regular expression syntax. */ 228 #define REG_EXTENDED 1 229 230 /* If this bit is set, then ignore case when matching. 231 If not set, then case is significant. */ 232 #define REG_ICASE (REG_EXTENDED << 1) 233 234 /* If this bit is set, then anchors do not match at newline 235 characters in the string. 236 If not set, then anchors do match at newlines. */ 237 #define REG_NEWLINE (REG_ICASE << 1) 238 239 /* If this bit is set, then report only success or fail in regexec. 240 If not set, then returns differ between not matching and errors. */ 241 #define REG_NOSUB (REG_NEWLINE << 1) 242 243 244 /* POSIX `eflags' bits (i.e., information for regexec). */ 245 246 /* If this bit is set, then the beginning-of-line operator doesn't match 247 the beginning of the string (presumably because it's not the 248 beginning of a line). 249 If not set, then the beginning-of-line operator does match the 250 beginning of the string. */ 251 #define REG_NOTBOL 1 252 253 /* Like REG_NOTBOL, except for the end-of-line. */ 254 #define REG_NOTEOL (1 << 1) 255 256 257 /* If any error codes are removed, changed, or added, update the 258 `re_error_msg' table in regex.c. */ 259 typedef enum 260 { 261 REG_NOERROR = 0, /* Success. */ 262 REG_NOMATCH, /* Didn't find a match (for regexec). */ 263 264 /* POSIX regcomp return error codes. (In the order listed in the 265 standard.) */ 266 REG_BADPAT, /* Invalid pattern. */ 267 REG_ECOLLATE, /* Not implemented. */ 268 REG_ECTYPE, /* Invalid character class name. */ 269 REG_EESCAPE, /* Trailing backslash. */ 270 REG_ESUBREG, /* Invalid back reference. */ 271 REG_EBRACK, /* Unmatched left bracket. */ 272 REG_EPAREN, /* Parenthesis imbalance. */ 273 REG_EBRACE, /* Unmatched \{. */ 274 REG_BADBR, /* Invalid contents of \{\}. */ 275 REG_ERANGE, /* Invalid range end. */ 276 REG_ESPACE, /* Ran out of memory. */ 277 REG_BADRPT, /* No preceding re for repetition op. */ 278 279 /* Error codes we've added. */ 280 REG_EEND, /* Premature end. */ 281 REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ 282 REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ 283 } reg_errcode_t; 284 285 286 /* This data structure represents a compiled pattern. Before calling 287 the pattern compiler, the fields `buffer', `allocated', `fastmap', 288 `translate', and `no_sub' can be set. After the pattern has been 289 compiled, the `re_nsub' field is available. All other fields are 290 private to the regex routines. */ 291 292 struct re_pattern_buffer 293 { 294 /* [[[begin pattern_buffer]]] */ 295 /* Space that holds the compiled pattern. It is declared as 296 `unsigned char *' because its elements are 297 sometimes used as array indexes. */ 298 unsigned char *buffer; 299 300 /* Number of bytes to which `buffer' points. */ 301 unsigned long allocated; 302 303 /* Number of bytes actually used in `buffer'. */ 304 unsigned long used; 305 306 /* Syntax setting with which the pattern was compiled. */ 307 reg_syntax_t syntax; 308 309 /* Pointer to a fastmap, if any, otherwise zero. re_search uses 310 the fastmap, if there is one, to skip over impossible 311 starting points for matches. */ 312 char *fastmap; 313 314 /* Either a translate table to apply to all characters before 315 comparing them, or zero for no translation. The translation 316 is applied to a pattern when it is compiled and to a string 317 when it is matched. */ 318 char *translate; 319 320 /* Number of subexpressions found by the compiler. */ 321 size_t re_nsub; 322 323 /* Zero if this pattern cannot match the empty string, one else. 324 Well, in truth it's used only in `re_search_2', to see 325 whether or not we should use the fastmap, so we don't set 326 this absolutely perfectly; see `re_compile_fastmap' (the 327 `duplicate' case). */ 328 unsigned can_be_null : 1; 329 330 /* If REGS_UNALLOCATED, allocate space in the `regs' structure 331 for `max (RE_NREGS, re_nsub + 1)' groups. 332 If REGS_REALLOCATE, reallocate space if necessary. 333 If REGS_FIXED, use what's there. */ 334 #define REGS_UNALLOCATED 0 335 #define REGS_REALLOCATE 1 336 #define REGS_FIXED 2 337 unsigned regs_allocated : 2; 338 339 /* Set to zero when `regex_compile' compiles a pattern; set to one 340 by `re_compile_fastmap' if it updates the fastmap. */ 341 unsigned fastmap_accurate : 1; 342 343 /* If set, `re_match_2' does not return information about 344 subexpressions. */ 345 unsigned no_sub : 1; 346 347 /* If set, a beginning-of-line anchor doesn't match at the 348 beginning of the string. */ 349 unsigned not_bol : 1; 350 351 /* Similarly for an end-of-line anchor. */ 352 unsigned not_eol : 1; 353 354 /* If true, an anchor at a newline matches. */ 355 unsigned newline_anchor : 1; 356 357 /* [[[end pattern_buffer]]] */ 358 }; 359 360 typedef struct re_pattern_buffer regex_t; 361 362 363 /* search.c (search_buffer) in Emacs needs this one opcode value. It is 364 defined both in `regex.c' and here. */ 365 #define RE_EXACTN_VALUE 1 366 367 /* Type for byte offsets within the string. POSIX mandates this. */ 368 typedef int regoff_t; 369 370 371 /* This is the structure we store register match data in. See 372 regex.texinfo for a full description of what registers match. */ 373 struct re_registers 374 { 375 unsigned num_regs; 376 regoff_t *start; 377 regoff_t *end; 378 }; 379 380 381 /* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, 382 `re_match_2' returns information about at least this many registers 383 the first time a `regs' structure is passed. */ 384 #ifndef RE_NREGS 385 #define RE_NREGS 30 386 #endif 387 388 389 /* POSIX specification for registers. Aside from the different names than 390 `re_registers', POSIX uses an array of structures, instead of a 391 structure of arrays. */ 392 typedef struct 393 { 394 regoff_t rm_so; /* Byte offset from string's start to substring's start. */ 395 regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ 396 } regmatch_t; 397 398 399 /* Declarations for routines. */ 400 401 /* To avoid duplicating every routine declaration -- once with a 402 prototype (if we are ANSI), and once without (if we aren't) -- we 403 use the following macro to declare argument types. This 404 unfortunately clutters up the declarations a bit, but I think it's 405 worth it. */ 406 407 /*#if __STDC__*/ 408 409 #define _RE_ARGS(args) args 410 411 /* #else */ /* not __STDC__ */ 412 413 /* #define _RE_ARGS(args) () */ 414 415 /* #endif */ /* not __STDC__ */ 416 417 /* Sets the current default syntax to SYNTAX, and return the old syntax. 418 You can also simply assign to the `re_syntax_options' variable. */ 419 NLM_EXTERN reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax)); 420 421 /* Compile the regular expression PATTERN, with length LENGTH 422 and syntax given by the global `re_syntax_options', into the buffer 423 BUFFER. Return NULL if successful, and an error string if not. */ 424 NLM_EXTERN const char *re_compile_pattern 425 _RE_ARGS ((const char *pattern, int length, 426 struct re_pattern_buffer *buffer)); 427 428 429 /* Compile a fastmap for the compiled pattern in BUFFER; used to 430 accelerate searches. Return 0 if successful and -2 if was an 431 internal error. */ 432 NLM_EXTERN int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer)); 433 434 435 /* Search in the string STRING (with length LENGTH) for the pattern 436 compiled into BUFFER. Start searching at position START, for RANGE 437 characters. Return the starting position of the match, -1 for no 438 match, or -2 for an internal error. Also return register 439 information in REGS (if REGS and BUFFER->no_sub are nonzero). */ 440 NLM_EXTERN int re_search 441 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, 442 int length, int start, int range, struct re_registers *regs)); 443 444 445 /* Like `re_search', but search in the concatenation of STRING1 and 446 STRING2. Also, stop searching at index START + STOP. */ 447 NLM_EXTERN int re_search_2 448 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, 449 int length1, const char *string2, int length2, 450 int start, int range, struct re_registers *regs, int stop)); 451 452 453 /* Like `re_search', but return how many characters in STRING the regexp 454 in BUFFER matched, starting at position START. */ 455 NLM_EXTERN int re_match 456 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, 457 int length, int start, struct re_registers *regs)); 458 459 460 /* Relates to `re_match' as `re_search_2' relates to `re_search'. */ 461 NLM_EXTERN int re_match_2 462 _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, 463 int length1, const char *string2, int length2, 464 int start, struct re_registers *regs, int stop)); 465 466 467 /* Set REGS to hold NUM_REGS registers, storing them in STARTS and 468 ENDS. Subsequent matches using BUFFER and REGS will use this memory 469 for recording register information. STARTS and ENDS must be 470 allocated with malloc, and must each be at least `NUM_REGS * sizeof 471 (regoff_t)' bytes long. 472 473 If NUM_REGS == 0, then subsequent matches should allocate their own 474 register data. 475 476 Unless this function is called, the first search or match using 477 PATTERN_BUFFER will allocate its own register data, without 478 freeing the old data. */ 479 NLM_EXTERN void re_set_registers 480 _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs, 481 unsigned num_regs, regoff_t *starts, regoff_t *ends)); 482 483 /* 4.2 bsd compatibility. */ 484 NLM_EXTERN char *re_comp _RE_ARGS ((const char *)); 485 NLM_EXTERN int re_exec _RE_ARGS ((const char *)); 486 487 /* POSIX compatibility. */ 488 NLM_EXTERN int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags)); 489 NLM_EXTERN int regexec 490 _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch, 491 regmatch_t pmatch[], int eflags)); 492 NLM_EXTERN size_t regerror 493 _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf, 494 size_t errbuf_size)); 495 NLM_EXTERN void regfree _RE_ARGS ((regex_t *preg)); 496 497 498 #ifdef __cplusplus 499 } 500 #endif 501 502 #undef NLM_EXTERN 503 #ifdef NLM_EXPORT 504 #define NLM_EXTERN NLM_EXPORT 505 #else 506 #define NLM_EXTERN 507 #endif 508 509 #endif /* not __REGEXP_LIBRARY_H__ */ 510 511 /* 512 Local variables: 513 make-backup-files: t 514 version-control: t 515 trim-versions-without-asking: nil 516 End: 517 */ 518 519